Quantitative Susceptibility Mapping for Differentiating Hydroxyapatite and Calcium Oxalate Breast Calcifications at 3T: A Phantom Study

Abstract

Objective. Breast microcalcifications cannot be differentiated by composition on mammography, and 70 to 80% of those biopsied prove benign. Type I calcifications are calcium oxalate (CaOx) and associate with benign processes; Type II are hydroxyapatite (HA) and correlate with malignancy. This work assesses whether quantitative susceptibility mapping (QSM) can distinguish the two at 3 T, and how much of the susceptibility present it recovers. Approach. HA and CaOx granules were co-embedded in alginate phantoms in three formulations: undoped, adipose-matched and fibroglandular-matched. Four replicates per condition were imaged at 0.70 and 0.86 mm isotropic resolution. The reconstruction pipeline, including the choice of background field removal, was verified stage by stage against a digital twin of known susceptibility. R2* was evaluated from the same acquisition. Main results. In undoped alginate, HA was detected in all eight measurements and CaOx in none. Recovery of known susceptibility was 18.2%, lost mainly to partial volume rather than reconstruction. Simulation places the susceptibility of the CaOx granules below 0.05 ppm relative to the gel. R2* was elevated at both particle types, 4.5 times more for HA. Fifteen of 48 measurements were affected by air retained within the porous granules, which can invert the sign of the recovered susceptibility. Significance. In undoped alginate, susceptibility detects HA alone while R2* is elevated at both, so together they distinguish HA, CaOx and no particle from one acquisition. In simulation, the 50% detection diameter rises from 0.40 mm at 0.70 mm resolution to 1.07 mm at 1.50 mm, against a clinical range of 0.1 to 1.0 mm. Trapped air is detectable from the sign of the local field without reconstruction.